Towards a deep learning model for hadronization

Hadronization is a complex quantum process whereby quarks and gluons become hadrons. The widely-used models of hadronization in event generators are based on physically-inspired phenomenological with many free parameters. We propose an alternative approach neural networks used instead. Deep generative highly flexible, differentiable, compatible Graphical Processing Unit (GPUs). make the first step towards data-driven machine learning-based model by replacing compont within Herwig generator (cluster model) Generative Adversarial Network (GAN). show that GAN capable reproducing kinematic properties cluster decays. Furthermore, we integrate this into to generate entire events can be compared output public simulator as well $e^+e^-$ data.